With over 1 million sufferers today, aphasia is a prevalent impairment after left-hemisphere stroke that affects the production or comprehension of spoken, written, or gestured speech. We contend, as have others, that effective prioritization and optimization of available treatment methods for aphasia requires an understanding of how the damaged system responds to different kinds of learning experiences, i.e., a theory of learning. This proposal is dedicated to furthering development of a theory of learning for aphasia rehabilitation based on powerful learning principles derived from basic psychological research (i.e., retrieval practice or RP principles). Our prior research provided an empirical foundation for the RP theory of learning by demonstrating its relevance for the treatment of naming impairment, a ubiquitous symptom experienced by people with aphasia (PWA) involving difficulty reliably naming common, everyday objects. In our prior work, we showed that treatment that provides `retrieval practice' (practice retrieving names for objects from long- term memory) conferred greater persistent benefit than a popular form of naming treatment that does not involve retrieval practice (errorless learning). In confirmation of a relate `spacing' principle, we also found treatment benefits were enhanced when repeated training trials for an item were spaced over time, as opposed to massed together. In this foundational work, we studied a relatively homogeneous group of PWA in that naming was the sole or primary area affected. In the current proposal we take important next steps to articulate the broader clinical and theoretical significance of our learning theory by studying how the RP principles impact naming impairment in aphasia more generally (Study 1); examining the principles' relevance for treating word comprehension deficits (i.e., problems reliably accessing the meanings of familiar words) in aphasia (Studies 1 & 2); and, by developing a theoretical account of how the RP principles impact word retrieval processes in production (Studies 3-5). Study 1 will examine how the RP principles impact naming impairment and the transfer of benefit from naming treatment to word comprehension performance in a heterogeneous and well-characterized sample of PWA. In individual differences analyses, we will examine variability in response to the principles as a function of breakdown in primary components of word processing and cognition, thereby providing benchmarks for how best to prioritize the different learning experiences examined given a PWA's profile of cognitive-linguistic deficits. Study 2 also focuses on comprehension, asking whether the RP principles influence the efficacy of receptive forms of treatment in PWA with severe word comprehension deficits. Study 3 examines conditions under which failed retrieval practice confers benefit. Study 4 investigates potential mechanisms to explain why effort enhances the benefits from successful retrieval practice. Finally, Study 5 will formalize the RP theory in a computational framework that will promote integration of existing models of lexical learning and use.